Russian legislation currently lacks specific civil liability norms for AI-related damages, relying instead on general Civil Code provisions. While the National AI Development Strategy and the AI Ethics Code emphasize human or organizational responsibility, these remain recommendations rather than binding legal requirements.
Mandatory insurance for AI-related harm is currently limited to experimental legal regimes, such as autonomous transport and telemedicine, effective since January 2025. Federal Law No. 243-FZ, enacted in July 2026, allows developers of national AI models to insure their liability, though the supporting regulations for these models only take effect on March 1, 2027.
The Russian insurance market lacks standardized AI liability products. Only SOGAZ offers individual programs, while companies like SberStrakhovanie and Ingosstrakh have limited or niche offerings. A study by the Actuarial Guild and DRT in October 2026 found that no insurers in Russia, Belarus, or Uzbekistan provide dedicated AI risk coverage, with 72% of firms having no plans to launch such products.
Insurers cite the lack of clear liability distribution between developers, integrators, and users, as well as the absence of standardized testing protocols and historical claims data, as primary barriers. Current cyber insurance policies typically cover external attacks, leaving internal AI operational errors, such as accidental data deletion, largely uncovered.
International practices also lack mandatory AI liability insurance. The EU AI Act focuses on monitoring and logging rather than insurance, while the US relies on a patchwork of state laws and court precedents. Recent cases, such as Moffatt v. Air Canada in 2024, confirm that companies remain liable for the actions of their AI agents, regardless of the technology's autonomy.
